Overview
Predictable Machines is building the next generation of verifiable AI systems—combining cutting-edge language models with formal verification, functional programming, and mathematical rigor. We are seeking a Verification-Focused AI Engineer who thrives at the intersection of AI capabilities and mathematical precision.
Ideal candidates are excited about LLMs but also focused on making them reliable, traceable, and mathematically sound; comfortable with functional programming paradigms; curious about formal methods; and capable of designing systems and workflows that integrate event-driven architectures, streaming systems, and verification pipelines.
Role context This role involves building verification systems that make AI trustworthy, not just impressive, and spans working with Kotlin and TypeScript to develop verification-first AI services and pipelines.
Responsibilities
- Build verification-first AI systems alongside senior engineers, focusing on Server-Sent Events architectures, streaming workflows, and real-time verification pipelines using Kotlin and TypeScript.
- Develop and integrate formal verification tools—work with SMT solvers, logical reasoning systems, and mathematical validation tools to ensure AI outputs are provably correct and traceable.
- Design streaming verification workflows that combine factual verification (web search), logical validation (formal methods), and mathematical checking (computational tools) into coherent, auditable pipelines.
- Implement TypeScript client libraries and UI components for real-time research steppers, verification progress visualization, and interactive audit trails with full type safety.
- Contribute to Docker-based tool ecosystem and maintain/extend containerized verification tools and deployment systems.
- Participate in verification methodology research and support enterprise integration patterns (authentication, multi-tenancy, API integrations) to embed verification capabilities in customer applications.
Background / Qualifications
- Strong foundation in Computer Science, Mathematics, or Engineering — degree preferred; exceptional self-taught candidates with demonstrated systems-building experience welcome.
- Proficiency in functional programming languages — experience with Kotlin, TypeScript, or Scala; comfortable with immutable data structures, composable functions, and type-safe architectures.
- Interest in mathematical reasoning and formal methods — curiosity about logic, proof systems, SMT solvers, or mathematical validation (coursework or personal projects welcome).
- Systems thinking mindset — experience with event-driven architectures, streaming systems, API design, or containerized applications; understands AI as part of larger, reliable systems.
- Collaborative engineering skills — comfortable with Git workflows, code reviews, and building production-quality software rather than just research prototypes.
Preferred / Bonus
- Formal methods exposure (theorem provers, model checking, constraint solving, or mathematical verification tools).
- LLM integration experience focused on reliability, evaluation, and systematic testing.
- Functional programming enthusiasm; experience with monads, type systems, or category theory.
- Enterprise software experience (authentication, multi-tenancy, observability, developer-friendly APIs).
- Interest in AI safety/explainability.
- Hands-on mentorship from experienced AI engineers and researchers; opportunity to work on impactful AI reliability projects; flexible, remote-first environment; potential transition to full-time roles; access to AI tools and learning resources.
What We Offer
- Opportunity to work on high-impact AI reliability projects with leadership exposure.
- Competitive salary based on experience.
- Career development, language courses, professional training, and certifications.
- Flexible, remote-first working environment.
- Access to state-of-the-art AI tools and learning resources.
Note: This refined description preserves the core content related to responsibilities, qualifications, and expectations while removing unrelated boilerplate and duplications. It is focused on the Verification-Focused AI Engineer role and keeps the core information intact without introducing new facts.